martiniii
Joined: January 2010
Posts: 2,112
Location:
|
Post by martiniii on Dec 15, 2010 13:54:37 GMT
I'm trying to gather screen shots for my review of Corpse Killer: Graveyard Edition and am not able to get the game working properly on SSF. First of all, SSF doesn't auto-detect the game's region. I got "Disc Unsuitable for this System" and had to manually switch it to "USA" to get the game to run.
Second, the sound is completely messed up. All I'm getting is what sounds like a cassette tape being played at 30x speed.
Neither of those matter since I'm just taking screen shots, but third, the FMV(which comprises 80% of the game) has taken an odd dive in quality. It's very pixelated and has unnaturally bright colors thrown in, rather like a video taken with a cell phone.
Any ideas? One thing I've noticed is that the red light in the corner is flashing VERY rapidly. I've tried toggling "Auto Field Skip" but that hasn't helped. Is it possible that SSF just has very low compatibility with this particular game, or with live action FMV in general?
|
|
|
|
Post by zyrobs on Dec 18, 2010 22:17:33 GMT
Linlhutz lists the game as working on version 0.10 Beta R1 but not on later ones, so try using that. First of all, SSF doesn't auto-detect the game's region. I got "Disc Unsuitable for this System" and had to manually switch it to "USA" to get the game to run. That's normal. The console doesn't switch regions automatically either. That's how it's supposed to look, but you didn't see it on your console because 1. oldschool CRT TVs blurred up the picture, and 2. standard composite connection blurred up the picture. If you'd play it on the console with an RGB cable and a high quality TV (possibly a hdtv with a good upscaler), it would look equally bad. For what its worth, you can enable bilinear filtering in SSF to smooth it out a little. |
|
|
|
Post by Kibbles on Dec 19, 2010 5:24:36 GMT
The main reason that Saturn FMVs look better on composite than they do on RGB is that the Saturn stores color information on Cinepak FMVs at a very low resolution and composite only contains very low resolution color information:-
So the main advantage to composite is that the low res color signal has been converted from a digital one into an analog one, and is being upscaled and converted using analog methods rather than digital ones, so no pixelation occurs and the end result is actually better looking!
Duck codec and MPEG FMVs look quite watchable through RGB though!.
|
|
|
|
Post by zyrobs on Dec 19, 2010 23:46:20 GMT
and composite only contains very low resolution color information: what Screen resolution has nothing to do with this, composite looks shit because of crosstalk between the brightness and color information. This can be reduced with better quality encoder/decoders, but no one gave a shit about that once s-video became a standard. It's the composite dot crawl effect that made checkerboard transparencies so prominent, and its part of the reason why shit quality FMVs looked okay on old consoles. If you used RGB cables and a decent TV, Saturn FMVs would've looked as much of a turd as they do on SSF. Alternatively, if your videocard has composite out (last cards to support that were the Radeon 4xxx ones), plug it into your TV and you'll get the same picture as how your Saturn would look. |
|
|
|
Post by Kibbles on Dec 20, 2010 5:34:09 GMT
and composite only contains very low resolution color information: what Screen resolution has nothing to do with this, composite looks shit because of crosstalk between the brightness and color information. This can be reduced with better quality encoder/decoders, but no one gave a shit about that once s-video became a standard. It's the composite dot crawl effect that made checkerboard transparencies so prominent, and its part of the reason why shit quality FMVs looked okay on old consoles. If you used RGB cables and a decent TV, Saturn FMVs would've looked as much of a turd as they do on SSF. Alternatively, if your videocard has composite out (last cards to support that were the Radeon 4xxx ones), plug it into your TV and you'll get the same picture as how your Saturn would look. I never said anything about 'screen' resolution. The bandwidth of the color channel in composite only has about 120 lines worth of information stored horizontally - which is then stretched and overlayed over the black and white image. Resolution has everything to do with it - The color signal inside a composite signal contains about 480x120 lines worth of information. Or whatever resolution you like by 120. Analog video formats for display on a CRT technically only have a line number on the direction they actually scan in, so on TV that's down. 'Resolution' isn't really contained as pixels or lines per-say on the horizontal plain, but rather voltage variations on the same line (which of course creates the pixels). So while it's not 100% accurate to say there are only 120 color values per line, that is approximately how much detail you are seeing. The reason you don't see the color distortion from the digital compression on composite is because the color information in a composite signal isn't actually enough to even show that it exists, and also because digital compression uses a similar method where the color information is stored at a lower resolution than luma information. Essentially, when you view a saturn's image through RGB, you are seeing the Chroma overlayed on the Luma as produced by the Saturn. When you view it on Composite or S-Video, you are seeing the same image produced by the saturn:- broken up back into Chroma and Luma then recombined by the analog decoder inside the television. This eliminated the color distortion effect because the color information has been blurred to hell by the low resolution chroma signal inside composite and S-video:- The pixels on the colors in the FMVs are blurred together but the Luma is left almost untouched. So, try understanding what is happening before waving your dick at me. Thanks. |
|
martiniii
Joined: January 2010
Posts: 2,112
Location:
|
Post by martiniii on Dec 22, 2010 23:04:08 GMT
Linlhutz lists the game as working on version 0.10 Beta R1 but not on later ones, so try using that. Thanks; I'll see if I can hunt that down. First of all, SSF doesn't auto-detect the game's region. I got "Disc Unsuitable for this System" and had to manually switch it to "USA" to get the game to run. That's normal. The console doesn't switch regions automatically either. Well, SSF always automatically switched regions before. I popped in a Japanese game, it played immediately and the "Area Code" menu was set to Japan. I popped in a USA game, it played immediately and the Area Code was set to "America, Canada, Brazil". I popped in a PAL game, it played immediately and the Area Code was set to "Europe, Australia, South Africa". Either SSF is designed to automatically detect region, or some glitch made SSF consistently switch to the correct region in a truly amazing coincidence. |
|
|
|
Post by zyrobs on Dec 22, 2010 23:54:55 GMT
OK, I'm not turning this into a dickfight. Your point, that the video pipeline with cvbs/svid is a mindfuck, that's true (original FMV (any kind of y/c resolution) -> RGB (saturn internals) -> YUV or YIQ (NTSC/PAL) -> back to RGB (pixels on your tv)). I didn't realize you were talking about color subsampling, since you worded it a little bit strange.
However, do note that:
This is wrong. I forgot which subsampling ntsc and pal uses, 4:2:2 or 4:2:0 (or the approximate for NTSC), but pal/ntsc uses approx 720 pixels by 480/576 lines. Halve that resolution for color, and you still get - worst case scenario with 4:1:0 - approx 360 by 240/288 color resolution. Considering that no Saturn games used double density or interlaced videos, the halved color data of pal/ntsc is already enough to carry the full color data for the low-res saturn display.
... and no, I'm not implying that s-video should be enough to get perfect picture. With S-Video you still have sync+luma on one cable, and yuv/yiq fucked up chroma on the other. Plenty of chance for crosstalk, plus the whole color subsampling thing. Meanwhile RGB has R, G, B and h/v sync all on separate cables, which obviously allows for better quality.
|
|
|
|
Post by Kibbles on Dec 23, 2010 21:18:16 GMT
As far as I'm aware the colour sub-sampling on composite and S-Video video is compressed only on the horizontal axsis, so a 4:1:0 ratio of compression on composite would yield a full number of vertical lines whilst the number of pixels horizontally is severley decreased - this is why colors only bleed out the side edges of objects on such video formats.
It probably should have read more like 120x480/120x576 than 480x120.
On composite and s-video, you never see colors bleed out the tops and bottoms of objects, only the sides- and analog image sharpening techniques only ever sharpen the sides of on-screen details!
The reason behind this method is so the color can be encoded into the picture on a per-line basis, the actual scanline resolution is kept full while only the X-Axsis' information is compressed.
The effect can be best seen on VHS, which uses a far higher compression level of sub-sampling than composite allows for, however the magnetic signal stored on the tape is simply a wave-form for a composite video signal - on VHS we are seeing about 40 odd lines of color information along the X-axsis! But full scanline resolution!
One of the main criticisms of Super-VHS video tape over VHS was that although the signals were now stored distinct from one and other on the tape, the chroma signal was still heavily compressed to levels way lower than the S-Video cable was able to allow for.
|
|
|
|
Post by zyrobs on Dec 25, 2010 0:25:05 GMT
I thought it was 4:2:0... I can't find any concrete info on this. Where did you get your info from?
And the reason why they employed this is because the human eye has less receptors for color than brightness, so there's no point in transmitting something we can't even see, and so by halving color information they could save precious bandwidth.
|
|
